<<

Zeitschrift fUr Kristallographie 175, 1-7 (1986) @ by R. Oldenbourg Verlag, Miinchen 1986

New family of silicate phases with the pollucite structure

Leticia M. Torres-Martinez and A. R. West

University of Aberdeen, Department of Chemistry, Meston Walk, Old Aberdeen AB9 2UE, Scotland, United Kingdom

Received: April 10, 1984; in final form: February 14, 1986

Pollucite structure I RbzMgSis012 type

Abstract. A new family of silicate phases, CszMSi5012: M = Be, Mg, Fe, Co, Ni, Zn, Cd; RbzMSi5012: M = Mg, Fe, (Co, Zn) has been prepared with the cubic pollucite structure, a in the range 13.3 to 13.8 A, Ia3d, Z = 8. The structure of one, RbzMgSi5012' has been refined to R = 10.8% using X-ray powder data. The structures are built of a 3D, [MSi50u]z- framework containing large, 12-coordinated cavities for the alkali cations. These structures are unusual in that a wide range of divalent cations have partially replaced Si4+ in the silicate anion framework.

Introduction Pollucite, CsAISiz06, is a framework alumino silicate containing large, 12- coordinated cavities suitable for occupation by Cs+ ions (Strunz, 1936; Naray-Szabo, 1938; Newnham, 1967). The unit cell is cubic with space group Ia3d, Z = 16. Similar alumino-silicate frameworks are present in analcite, NaAISiz06 . HzO and leucite, KAISiz06 (Taylor, 1938; Beattie, 1954; Knowles et aI., 1965; Wyart, 1941; Deer et aI., 1963; Wyckoff, 1968), although leucite is tetragonal below 6050 (Wyart, 1941; Deer et aI., 1963; Wyckoff, 1968). . Structural studies on these phases are complicated by the occurrence of twinning in tetragonalleucite (Mazzi et aI., 1976) and by Si, Al disorder in the alumino silicate framework. Various analogues of pollucite and leucite, containing other trivalent ions instead of AIH have been reported: Fe- pollucite, CsFeSiz06 (Kopp et aI., 1963; Kume and Koizumi, 1965), Fe- 2 L. M. Torres-Martinez and A. R. West leucite, KFeSi206, tetragonal RbFeSi206 (Hirao et aI., 1976) and tetragonal RbAISi206 (Hirao et aI., 1976; Barrer et aI., 1953; Taylor and Henderson, 1968). Recently, we have synthesised Cs2BeSis012 (Torres- Martinez et aI., 1984) with the pollucite structure. It seemed likely that a variety of other divalent ions should also enter the pollucite framework; this was found to be the case and these results are reported here.

Experimental Starting materials were Si02 (high purity, crushed ), CS2C03 (Aldrich, 99.9% pure), Rb2C03 (Fluka, 98% pure), MgO (Hopkin and Williams, Analar), Fe(COOh, CoC03, NiC03, ZnO, CdO (all BDH, re- agent grade). Mixtures in the stoichiometric ratio, (Mi C03 :M2+0: 5 Si02 with M+ = Rb, Cs; M2+ = Mg, Zn etc. were prepared in", 10 g quantities by mixing into a paste with acetone, drying and firing in Pt crucibles or foil boats, initially at 600 to 700 °C to expel C02 and finally at 900 to 1200 °C for 1 to 3 days, depending on composition, to complete the reac- tion. Some samples were melted, quenched to form a glass, and sub- sequently crystallised at temperatures in the range 900 to 1200°C. The products were identified by X-ray powder diffraction using a Hiigg Guinier focusing camera, CuKal radiation. For accurate d-spacing measurements, analar grade KCI, a = 6.2931 A, was added as an internal standard and the films were measured using a Cooksley microdensitometer. The powder patterns were indexed by comparison with that of Cs2BeSis012 (Torres- Martinez et aI., 1984) and unit-cell dimensions obtained by least-squares refinement. Powder intensities suitable for a structure determination were obtained using a Philips diffractometer, CuKa radiation, with a slow scan speed of 1-° min - 1. Intensities, measured from the diffractometer chart by using the method of counting squares to give peak areas, were corrected for Lp factors and multiplicities in the usual way. Structure refinement was carried out with a least-squares program using Fhklvalues. Melting points were determined by placing powdered samples, wrapped in Pt foil, in a vertical tube furnace whose temperature was controlled and measured to ::!::5°C. The samples were heated for short times, '" 1-h, quenched into Hg and their X-ray powder patterns recorded. The liquids that formed on melting most of the new phases readily formed glasses on cooling to room temperature. It was therefore an easy matter to determine whether or not melting had occurred from the presence (not melted) or absence (melted) of lines in the X-ray powder patterns of the quenched samples. Several of the phases were found to exhibit polymorphism. This was shown by heating samples at selected temperatures followed by either quenching or slowly cooling to room temperature. Silicate phases with the pollucite structure 3

Results and discussion

Ten new silicate phases of general formula MiM2+Sis012 with the pollucite structure have been synthesised. They were prepared either by solid state reaction or by crystallisation of glass or by both methods. All are thermodynamically stable and melt congruently at high temperatures. Some are stable only at high temperatures and transform to other polymorphs on cooling. Crystallographic data, melting points, conditions of synthesis and polymorphism of the new phases, together with data for Cs2BeSis012 (Torres-Martinez et aI., 1984) are given in Table 1. The new, pollucite-like phases, labelled IXphases, are cubic with space group Ia3d. For two of them, Cs2CoSis012 and Rb2MgSis012, the space group was confirmed using selected area electron diffraction. Indexed X- ray powder data for Rb2MgSis012 are given in Table 2; data for the remainder have been submitted to the JCPDS X-ray powder diffraction file. The structure of one of the new phases, Rb2MgSis012, was confirmed by least-squares refinement of the observed structure factors, Fhk1,using as starting coordinates, the refined parameters of Cs2BeSis012 (Torres- Martinez et aI., 1984). Initially, 18 uniquely indexed reflections were used in the refinement and an R value of 10.1 % obtained. Subsequently, an additional 4 overlapping reflections were included. Assignment of in- tensities to the component peaks was made using the calculated structure factors as a guide. After the final refinement using a total of 27 reflections, an R value of 10.8% was obtained. For the final refinement, all relevant positional coordinates and the isotropic temperature factors were allowed to refine independently. The relevant data are given in Table 2 with final atomic coordinates in Table 3 and a selection of bond distances and angles in Table 4. The structure of Rb2MgSis012 is essentially the same as that of pollucite, CsAlSi206 (Strunz, 1936; Naray-Szab6, 1938; Newnham, 1967) and Cs2BeSis012 (Torres-Martinez et aI., 1984). It may be described as a 3D magnesiosilicate framework built of corner-sharing (Mg, Si)04 tetrahedra with the large Rb+ ions in 12-coordinate cavities. There was no evidence of ordering of Mg and Si over the 48 (g) sites. A comparison of Cs2BeSis012 and Rb2MgSis012 shows that, while Rb2MgSis012 has the larger unit-cell dimension, six of the Rb-O bonds in it are shorter than the corresponding six Cs - 0 bonds in Cs2BeSis012' There appear to be two factors involved in this effect. First, introduction of Mg in place of Be in the anion framework causes a small, average increase in the (M2+ ,Si4+)-O bond lengths. Second, the anion framework becomes slightly distorted so as to allow six shorter Rb - 0 bonds. The temperature factors given for Rb2MgSis012, Table 3, are all rather large. This may be variously attributed to (a) the limited data set used (b) Table 1. Properties of new polucite-like phases

Phase Conditions of synthesis (IXpolymorph) Other polymorphs, a(A) Melting point, conditions of synthesis °C

Cs2BeSis012 crystallisation of glass 900 °C 13.406(1 ) 1420:t 20 or solid state reaction, 1000-1200°C y, solid state reaction Cs2MgSis012 crystallisation of glass 900 °C 13.695(1) 1450::!::20 or solid state reaction, 1400°C at 1000-1200°C Cs2FeSis012 solid state reaction, 1100°C 13.834(2) 1420::!::30 y(or D), solid state reaction Cs2CoSis012 crystallisation of glass 1100 ° C 13.672(1) 1340 :t 20 1000-1200°C Cs2NiSis012 1300°C, quench D(or y), 1200°C 13.654(1) 1325 :t 50

Cs2ZnSis012 crystallisation of glass 900 °C; D(ory),1100-1300°C 13.662(2) 1400:t 20 1380 °C, quench Cs2CdSis012 solid state reaction, 1000°C 13. 791 (2) 1090:t 20 Rb2MgSis012 1350°C, solid state reaction fI, non-cubic 1100-1300°C 13.530(1) 1360 :t 40

Rb2FeSis012 crystallisation of glass 1000 °C 13.650( 4) 1070:t 20 Rb2CoSis012* 1250 ° C, quench * fI, non-cubic 1100-1200°C 1260:t 10 1325 :t 20 Rb2ZnSis012 * 1300°C, quench* fI, non-cubic 1100-1200°C It was not possible to obtain a pure IXpolymorph; partial transformation to fI occurred, even on rapid quenching * It was not possible to obtain a pure IXpolymorph free from one or two weak extra lines which could be associated with either y or fI * polymorphs Silicatephases with the pollucite structure 5

Table 2. X-ray powder data for RbzMgSis012' a = 13.530:!:: 0.001 A, S.G. la3d hkl dobs(A) dcalc(A) lobs Fobs Fca1c

022 4.805 4.784 1.7 6.3 9.9 123 3.623 3.616 24.7 16.2 18.5 004 3.387 3.383 100 99.3 96.6 024 3.029 3.026 3.8 11.0 9.7 233 2.887 2.885 38.0 36.4 35.4 134 2.656 2.654 4.9 10.1 10.8 125 2.472 2.470 1.4 6.0 6.1 044 2.3921 2.3918 18.6 44.5 47.1 116 4.8 4.8 2.1949 2.1949 7.7 235 ) 15.4 16.5 136 1.9945 1.9949 3.9 12.5 12.6 444 1.9535 1.9529 2.6 25.7 20.6 345 1.9133 1.9135 0.9 6.3 4.6 046 1.8772 1.8763 1.0 9.9 12.9 127} 14.4 15.0 255 1.8411 1.8412 8.7 8.7 9.4 336 19.2 20.8 23.2 27.8 237 1.7180 1.7184 9.5 156 ) 3.9 4.6 008 1.6910 1.6913 3.0 38.3 35.8 147 1.6656 1.6655 2.5 12.5 10.5 028 1.6409 1.6408 1.9 15.7 11.5 356 1.6163 1.6172 2.0 11.6 7.8 138 12.2 11.4 1.5729 1.5729 2.2 347 ) 3.1 2.8 257 1.5317 1.5320 1.6 11.2 11.8 048 1.5129 1.5127 2.3 19.3 18.6

Table 3. Atomic parameters for RbzMgSis012

Atom Position x y Z Bisa

0 96 h 0.100:!:: 0.001 0.127 :!::0.002 0.717:!:: 0.001 4.0 Si, Mg 48 g 0.125 0.665 :!::0.001 0.585 :!::0.001 2.3 Rb 16 b 0.125 0.125 0.125 6.0

Table 4. Some bond distances and angles

(Mg,Si)-O 1.610:!:: 0.022 A (2 x) 1.628:!:: 0.019 A (2 x) Rb-O 3.598 :!::0.018 A (6 x) 3.291 :!::0.018 A (6 x) 0 O-(Mg,Si)-O 104.7 :!::1.1 118.7 :!::1r (2 x ) 106.9 :!::1.1 0 (2 x ) 101.8 :!::lr 6 L. M. Torres-Martinez and A. R. West disorder in positions, associated with different sizes of the randomly arranged Mg04, Si04 tetrahedra (c) disorder in Mg, Si positions and (d) the rather long Rb - bond lengths and therefore, large vibration ° amplitudes for Rb. Further evidence of disorder in the structure is indicated by the distorted average shape of the (Mg, Si)04 tetrahedra in which the bond angles vary from 102 to 119°. In Cs1BeSis012, by comparison, the bond angles were all between 104 and 1130. In summary, therefore, it is considered that the atomic coordinates given in Table 3 represent an average structure and that on a microscopic scale, considerable structural distortion and disorder is likely to be present. Variations in lattice parameters, a, between the different phases reported in Table 1 are in approximate accord with the relative sizes of the component M+, M2+ ions. Thus, for the Cs+ phases the lattice parameter of the transition metal ion-containing phases gradually decreases from M = Fe to Zn; the sizes of the Mg, Zn-containing phases are approximately equal. Further, for a given M2+ ion, the lattice parameter of a Cs+ phase is larger than that of the Rb + phase. However, all of the lattice parameter values given in Table 1 fall within a span of only 3% which is much '" smaller than the spread in the radii of either the M+ or M2+ ions that are present. This indicates that the lattice parameter value is controlled primarily by the size of the silicate anion framework. The latter is modified to a small degree by the various divalent ions that substitute into the framework and by the different alkali cations that occupy the large, 12- coordinate cavities. Several of the new phases are polymorphic, as given in Table 1. The polymorphs fall into two main categories. The ')', 15polymorphs are cubic, like the (f.polymorphs, but show extra weak lines in their X-ray powder patterns. The 15polymorphs have a primitive cubic space group with no systematic absences, as shown for the structurally related germanate phases, Cs1ZnGeS012, Cs1CoGeS012 and Rb1MgGes012 (Torres-Martinez et aI., 1984). The')' polymorphs are primitive cubic but have systematic absences associated with a single a glide plane and probable space group Pa 3 (Torres- Martinez et al.). In some of the phases it was not possible to distinguish from powder data alone between 15and')' polymorphs. The lattice constants of the 15,')' polymorphs were not determined but, in each case, appeared to be very similar to those of the parent (f.polymorphs. The 11polymorphs are non-cubic and are characterised by split lines in their X-ray powder patterns in comparison with the (f.polymorphs. The 11 polymorphs are not simply tetragonal, leucite-like phases (Torres-Martinez et al.) although the related phase, Rb1BeSis012 (labelled e) is leucite-like (Torres-Martinez et ai., 1984). The 11polymorphs appear to have symmetry lower than tetragonal (Torres-Martinez et al.). It also appears to be generally the case that in those phases that show polymorphism, the (f.,pollucite-like polymorph is stable at high tempera- Silicate phases with the pollucite structure 7 tures and the y, t5and Yfpolymorphs are stable at lower temperatures. The transitions between IXand y, t5or Yfpolymorphs were generally found to be reversible.

Acknowledgements. One of us (L.M.T.-M.) wishes to thank CONACYT, Mexico, for a scholarship. Dr. J. A. Gard checked the space groups of two of the new phases using electron diffraction. Dr. R. A. Howie provided general comments and advice on the computer programs. The computing was carried out on the University Honeywell com- puter.

References Barrer, R. M., Baynham, J. W., McCallum, N.: Hydrothermal chemistry of silicates. Part V. Compounds structurally related to analcite. J. Chern. Soc. 21, 4035-4041 (1953) Beattie, 1. R.: The structure of analcite and ion-exchanged forms of analcite. Acta Crystallogr. 7, 357-359 (1954) Deer, W. A., Howie, R. A., Zussman, J.: Rock Forming Minerals, Vol. 4, p. 276, 338. New York: Longmans 1963 Rirao, K., Soga, N., Kunugi, M.: Thermal expansion and structure of leucite-type compounds. J. Phys. Chern. 80,1612-1616 (1976) Knowles, C. R., Rinaldi, F. F., Smith, J. V.: Refinement of the of . Indian Miner. 6,127-140 (1965) Kopp, O. C., Harris, L. A., Clarke, G. W., Yakel, H. L.: A hydrothermally synthesised analog of pollucite - its structure and significance. Am. Mineral. 48, 100-1 09 (1963) Kume,S., Koizumi: Syntheticpollucitesin the systemCSzO. Alz03 . 4SiOz - CSzO . FeZ03 . 4SiOz-HzO - their phase relationship and physical properties. Am. Mineral. 50, 587- 592 (1965) Mazzi, F., Galli, E., Gottardi, G.: The crystal structure of tetragonal leucite. Am. Mineral. 61,108-115 (1976) Naray-Szab6, St. v.: Die Struktur des Pollucits CsAlSiz06 . xHzO. Z. Kristallogr. 99, 277- 282 (1938) Newnham, R. E.: Crystal structure and optical properties of pollucite. Am. Mineral. 52, 1515-1518 (1967) Strunz, H.: Die chemische Zusammensetzung von Pollucit. Z. Kristallogr. 95, 1-18 (1936) Taylor,D., Henderson, C. M. B.: The thermal expansion ofthe leucite group of minerals. Am. Mineral. 53,1476-1489 (1968) Taylor, W. H.: Note on the structures of analcite and pollucite. Z. Kristallogr. 99, 283- 291 (1938) Torres-Martinez, L. M., Gard, J. A., Howie, R. A., West, A. R.: Synthesis ofCszBeSisO 12 with a pollucite structure. J. Solid State Chern. 51,100-103 (1984) Torres-Martinez, L. M., Gard, J. A., West, A. R.: Synthesis and structure ofa new family of phases AzMGes012: A = Rb, Cs; M = Be, Mg, Co, Zn. J. Solid State Chern. 53, 354- 359 (1984) Torres-Martinez, L. M., Gard, J. A., West, A. R.: unpublished Torres-Martinez, L. M., West, A. R.: Synthesis of RbzBeSis012 with a leucite structure. J. Mater. Sci. LeU. 3, 1093-1094 (1984) Wyart, J.: Structure atomique de la leucite. Compt. Rend. Acad. Sci. Paris 212, 356- 358 (1941) Wyckoff, R. G.: Crystal Structures, Vol. 4, p. 400. New York: Wiley 1968